Can I run Llama 3.1 70B on Apple M3 Max (64GB)?
Yes: Llama 3.1 70B runs on Apple M3 Max (64GB) with IQ4_XS at roughly 6 tokens/s. Estimated memory 39.5 GiB at 8,192 tokens of context.
With IQ4_XS you can push the context to about 19,456 tokens before memory runs out (F16 KV cache).
Memory by quantisation (8,192-token context)
| Quantisation | Memory | Verdict | Est. speed |
|---|---|---|---|
| F16Unquantized reference | 136.0 GiB | Does not fit | — |
| Q8_0Near-lossless | 74.4 GiB | Needs CPU offload | — |
| Q6_KVery close to F16 | 58.5 GiB | Needs CPU offload | — |
| Q5_K_MGreat quality/size | 51.3 GiB | Needs CPU offload | — |
| Q4_K_MBest default for most | 44.4 GiB | Tight | ~6 tok/s |
| IQ4_XSSmaller 4-bit (imatrix) | 39.5 GiB | Fits | ~6 tok/s |
| Q3_K_MNoticeable loss | 36.7 GiB | Fits | ~7 tok/s |
| Q2_KLast resort; big loss | 32.1 GiB | Fits | ~8 tok/s |
Apple M3 Max (64GB) offers about 64 GiB of memory (75% usable for a model) at 400 GB/s. Speed is a bandwidth-bound estimate of single-stream generation at 65% efficiency; real results vary by runtime, drivers and settings, so measure on your machine.
How this is calculated
Weights ≈ parameters × bits per weight ÷ 8. KV cache ≈ 2 × 80 layers × 8 KV heads × 128 head dim × 2 bytes per token. Decode speed ≈ efficiency × memory bandwidth ÷ bytes read per token. Learn the maths in VRAM math and memory bandwidth, or try the VRAM calculator.
Llama 3.1 70B on other hardware
Similar models on Apple M3 Max (64GB)
Running slowly or out of memory? See CUDA out of memory and the quantisation lesson.